Inorg. Chem. 2009, 48, 1114-1121
Model Ziegler-Type Hydrogenation Catalyst Precursors, [(1,5-COD)M(µ-O2C8H15)]2 (M ) Ir and Rh): Synthesis, Characterization, and Demonstration of Catalytic Activity En Route to Identifying the True Industrial Hydrogenation Catalysts William M. Alley,† Chase W. Girard,† Saim O¨zkar,‡ and Richard G. Finke*,† Department of Chemistry, Colorado State UniVersity, Fort Collins, Colorado 80523, and Department of Chemistry, Middle East Technical UniVersity, 06531 Ankara, Turkey Received October 11, 2008
The compounds [(1,5-COD)M(µ-O2C8H15)]2 (COD ) cyclooctadiene, M ) Ir (1) or Rh (2), O2C8H15 ) 2-ethylhexanoate) were synthesized by addition of Bu3NH(2-ethylhexanoate) or Na(2-ethylhexanoate) to acetone suspensions of [(1,5-COD)Ir(µ-Cl)]2 or [(1,5-COD)Rh(µ-Cl)]2, respectively. The synthesis of such well-defined second and third row model precursors is key to determining the true nature of commercial Ziegler-type hydrogenation catalysts (i.e., catalysts made from the combination of a non-zerovalent, group 8-10 transition metal precatalyst and a trialkylaluminum cocatalyst), an unsolved, ∼40 year old problem. The characterizations of 1 and 2 were accomplished by elemental analysis, melting point, FAB-MS, FT-IR, UV-vis, NMR spectroscopy, and single crystal X-ray diffraction. The complexes, C32H54Ir2O4 and C32H54O4Rh2, are isostructural: monoclinic, P21/n, Z ) 4. The lattice constants for 1 are a ) 15.7748(5) Å, b ) 9.8962(3) Å, c ) 20.8847(7) Å, β ) 108.408(2)°. The lattice constants for 2 are a ) 15.7608(4) Å, b ) 9.9032(3) Å, c ) 20.8259(5) Å, β ) 108.527(1)°. Complexes 1 and 2 are dimeric, bridged by the 2-ethylhexanoates, and with one 1,5-COD ligand bound to each metal. The formally 16 electron metal atoms are in square ligand planes with dihedral angles between the planes of 56.5° for 1 and 58.1° for 2. The M-M distances of 3.2776(2) and 3.3390(4) Å for 1 and 2, respectively, fall in the range of similar structures thought to have some M-M interaction despite the lack of a formal M-M bond. Demonstration that active Ziegler-type hydrogenation catalysts are made when 1 or 2 combine with AlEt3 is provided, results that open the door to the use of 1 and 2 as well-defined third and second row congeners, respectively, of Ziegler-type hydrogenation catalysts. These compounds have proven important in addressing the previously unsolved problem of the true nature of the catalyst in industrial Ziegler-type hydrogenation catalyst systems; their high yield synthesis and unequivocal characterization reported herein are the necessary first steps of that work.
Introduction The selective catalytic hydrogenation of unsaturated sites in polymers, such as styrenic block copolymers, is an important industrial process used to improve the stability of the polymer toward both thermal and oxidative degradation by autoxidation processes.1-3 According to one estimate, annual worldwide production of hydrogenated styrenic block * To whom correspondence should be addressed. E-mail: rfinke@ lamar.colostate.edu. † Colorado State University. ‡ Middle East Technical University. (1) McManus, N. T.; Rempel, G. L. J.M.S. ReV. Macromol. Chem. Phys. 1995, 35, 239–285. (2) Johnson, K. A. Polym. Prepr. 2000, 41, 1525–1526.
1114 Inorganic Chemistry, Vol. 48, No. 3, 2009
copolymers probably exceeds 1.7 × 105 metric tons.4 An industrially important family of catalysts developed during the late 1960s and early 1970s for the purpose of polymer hydrogenation is Ziegler-type hydrogenation catalysts, defined as those made from a non-zerovalent, group 8-10 transition metal precatalyst and a trialkylaluminum (e.g., (3) Weiner, H.; Trovarelli, A.; Finke, R. G. J. Mol. Catal. A: Chem. 2003, 191, 217–252. (4) Johnson, K. A.; Houston, TX.2 Personal communication, 2008. The complete statement was: “In their March 2005 study of Syrenic Thermoplastic Elastomers (525.8600 A), SRI estimated the global 2004 SBC consumption at 1,293,000 tonnes with HSBC at 10% of that. Since then we believe demand growth rates have been of the order of 10% pa which would imply a 2007 demand in the region of 170,000 tonnes for HSBC.”
10.1021/ic801933b CCC: $40.75
2009 American Chemical Society Published on Web 01/07/2009
Model Ziegler-Type Hydrogenation Catalyst Precursors
triethylaluminum) cocatalyst. As such, Ziegler-type hydrogenation catalysts are used primarily for polymer hydrogenation, not polymerization,1,2,5-10 and should not be confused with Ziegler-Natta polymerization catalysts. Surprisingly little fundamental information about Zieglertype hydrogenation catalysts exists despite their ∼40 year history of industrial application.2,11,12 The single most important unanswered question is what is the true nature of Ziegler-type hydrogenation catalysts: are they single metal “homogeneous” or multiple metal “heterogeneous” catalysts?13-15 The study of typically ill-defined industrial catalyst precursors alone has not led to a clear understanding of the true nature of these important industrial catalysts.2,8,9,16 Preparing well characterized, second and third row precursors, that yield model catalysts amenable to characterization by modern methods is the crucial first step in attaining new insights into the true nature of Ziegler-type hydrogenation catalysts.17 Requirements for an ideal precursor are that it (i) fits the above definition of a Ziegler-type hydrogenation catalyst (i.e., that it is composed of a group 8-10 transition metal salt); (ii) has an anion such as 2-ethylhexanoate that is representative of those commonly employed industrially;2 (iii) forms a catalytically active species for olefin hydrogenation upon combination with a typical alkylaluminum reagent such as AlEt3; and (iv) is readily available, preferably inexpensive and easy to prepare reproducibly in a well characterized, pure form. Additional requirements for an ideal precursor are that it: (v) be soluble in solvents commonly used for olefin hydrogenation, such as cyclohexane for example; (vi) have an auxiliary ligand, such as 1,5-COD, which can be used as an analytical handle (e.g., to monitor conversion of the precatalyst to the catalyst);18 (vii) employ a third row transition metal to allow transmission electron microscopy (TEM) to reliably image any polymetallic
(5) Lapporte, S. J. Preparation of Complex Organic Metallic Hydrogenation Catalysts and Their Use. U.S. Patent 3,205,278, Sep 7, 1965. (6) Cannell, L.; Magoon, E. F.; Raley, J. H. Olefin Oligomerization. U.S. Patent 3,424,815, Jan 28, 1969. (7) Wald, M. M.; Quam, M. E. Selectively Hydrogenated Block Copolymers. U.S. Patent 3,700,633, Oct 24, 1972. (8) Sloan, M. F.; Matlack, A. S.; Breslow, D. S. J. Am. Chem. Soc. 1963, 85, 4014–4018. (9) Kroll, W. J. Catal. 1969, 15, 281–288. (10) Lapporte, S. J. Ann. N. Y. Acad. Sci. 1969, 158, 510–525. (11) Barrault, J.; Blanchard, M.; Derouault, A.; Ksibi, M.; Zaki, M. I. J. Mol. Catal. 1994, 93, 289–304. (12) Pasynkiewicz, S.; Pietrzykowski, A.; Dowbor, K. J. Organomet. Chem. 1974, 78, 55–59. (13) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles and Applications of Organotransition Metal Chemistry; University Science Books: Mill Valley, CA, 1987. (14) Widegren, J. A.; Finke, R. G. J. Mol. Catal. A: Chem. 2003, 198, 317–341. (15) Schwartz, J. Acc. Chem. Res. 1985, 18, 302–308. (16) Falk, J. J. Polym. Sci., Part A-1 1971, 9, 2617–2623. (17) Alley, W. M.; Kayiran, I.; Wang, Q.; Frenkel, A.; Long, L.; Yang, ¨ zkar, S.; Johnson, K. A.; Finke, J. C.; Menard, L. D.; Nuzzo, R. G.; O R. G.; Ziegler-type Hydrogenation Catalysts Made From [(1,5COD)Ir(µ-O2C8H15)]2 and AlEt3 to be submitted for publication. (18) (a) Watzky, M. A.; Finke, R. G. J. Am. Chem. Soc. 1997, 119, 10382– ¨ zkar, S.; Finke, R. G. Langmuir 2003, 19, 6247–6260. 10400. (b) O
clusters that may be formed;19 (viii) form the same fundamental type (i.e., homogeneous or heterogeneous) of catalyst made by industrially used precatalysts, and therefore, (ix) yield new insights into the Ziegler-type hydrogenation catalysts central to industrial polymer hydrogenation. A literature search revealed that complexes of the form [(1,5COD)M(µ-O2C8H15)]2 (COD ) cyclooctadiene, M ) Ir or Rh, O2C8H15 ) 2-ethylhexanoate) are perhaps ideal, previously unexploited, candidates for the desired precatalyst.20-37 In addition, [(1,5-COD)M(µ-O2C8H15)]2 (M ) Ir or Rh) could prove to be of importance in a variety of other industrial applications.38 Herein we describe the synthesis, characterization, structural determination, and catalytic hydrogenation activity following the addition of AlEt3 of complexes [(1,5-COD)Ir(µO2C8H15)]2 (1) and [(1,5-COD)Rh(µ-O2C8H15)]2 (2). The methods used for compositional and structural characterization are elemental analysis, mass spectrometry (MS), single crystal X-ray diffractometry, infrared (IR), UV-visible electronic absorption, and nuclear magnetic resonance (NMR) spectroscopy. In work to be reported separately,17 we have used 1 and 2 as Ziegler-type hydrogenation catalyst precursors to investigate the problem of the true nature of industrial Ziegler-type polymer hydrogenation catalysts, a perplexing problem in the “is it homogeneous or heterogeneous catalysts?” area.14 That subsequent work hinges on the high yield synthesis and unequivocal characterization of the [(1,5COD)M(µ-O2C8H15)]2 (M ) Ir or Rh) precursors described herein. Results and Discussion Synthesis of [(1,5-COD)Ir(µ-O2C8H15)]2 (1) and [(1,5COD)Rh(µ-O2C8H15)]2 (2). The syntheses of 1 and 2, Scheme 1, were modeled after earlier syntheses of similar complexes which used [(diene)M(µ-Cl)]2 (diene ) 1,5-COD or norbornadiene (NBD), M ) Ir or Rh), and Na+, K+, or Ag+ carboxylate salts as starting materials, Supporting Information, Table S1. Control experiments were performed with longer reaction times while following the formation of (19) (a) Jaska, C. A.; Manners, I. J. Am. Chem. Soc. 2004, 126, 1334– 1335. (b) Jaska, C. A.; Manners, I. J. Am. Chem. Soc. 2004, 126, 9776–9785. (c) Hagen, C. M.; Widegren, J. A.; Maitlis, P. M.; Finke, R. G. J. Am. Chem. Soc. 2005, 127, 4423–4432. (d) Hagen, C.; VieillePetit, L.; Laurenczy, G.; Su¨ss-Fink, G.; Finke, R. G. Organometallics 2005, 24, 1819–1831. (e) Williams, D. B.; Carter, C. B. Transmission Electron Microscopy; Plenum Press: New York, 1996. (20) The similarity of [(1,5-COD)M(µ-O2C8H15)]2 compounds (M ) Ir or Rh, O2C8H15 ) 2-ethylhexanoate) to industrial catalyst precursors such as Ni(2-ethylhexanoate)2 made it very likely that they would form active Ziegler-type hydrogenation catalysts upon addition of AlEt3 (as demonstrated herein). Additionally, complexes of the type [(1,5COD)M(µ-carboxylato)]2 (COD ) cyclooctadiene, M ) Ir or Rh) were expected to be, and are, soluble in the same solvents as Ni(2ethylhexanoate)2 and other industrial catalyst precursors. Gas chromatographic (GC) analysis of cyclooctane (i.e., hydrogenated 1,5COD) has proven to be an important analytical handle in past nanocluster syntheses from (1,5-COD)M+ (M ) Rh, Ir) precursors.18 It seemed especially promising to be able to use catalysts formed from the heavier, 3rd row transition metal Ir complex, as they would more likely be amenable for use with TEM to look for the presence of metal particles than catalysts made from lighter Co and Ni counterparts. The ability to compare the Ir compound to its Rh analogue, as well as to industrial Ni and Co catalysts, is an additional advantage of the present approach and precursors.
Inorganic Chemistry, Vol. 48, No. 3, 2009
1115
Alley et al. a
Scheme 1. Stoichiometry and Associated Reaction Conditions for the Synthesis Of 1 (Top) and 2 (Bottom)
a Characterization for both compounds was accomplished using by elemental analysis, FAB-MS, single crystal X-ray diffraction, FT-IR, UV-Vis, and 1H and 13C spectroscopy, vide infra.
1 and 2 directly with 1H NMR. The reactions proceeded rapidly at room temperature (complete reaction in e10 min). Additionally, identical product yields were obtained if the reaction time was 10 min or 9 days. The synthesis of 1 was also accomplished using either the Na+ or Ag+ salt of 2-ethylhexanoate. However use of the Na+ salt requires an additional filtration step while use of the Ag+ salt gave only limited amounts of 1. Limited product formation was also observed in the attempted use of Bu3NHO2C8H15 to prepare 2. The necessity of using Na+ for the synthesis of 2 could result from different driving force requirements in the syntheses of 1 and 2. The hexane/H2O extraction step, shown in Scheme 1, was performed outside the drybox, with the exclusion of air accomplished using Ar pressure and cannula techniques, and rigorously degassed hexane and water. Compounds 1 and 2 were easily crystallized from acetone with slow cooling; performing the extractions thoroughly to (21) Although [(1,5-COD)M(µ-O2C8H15)]2 (M ) Ir or Rh) are not commercially available, their ready synthesis also appeared promising. Chatt and Venanzi described the original synthesis of the acetate compound [(1,5-COD)Rh(µ-O2CCH3)]2 from [(1,5-COD)Rh(µ-Cl)]2 and potassium acetate.26 The synthesis of other [(diene)Rh(µ-O2CR)]2 complexes have been reported using similar approaches, including those where the R group is CH2Cl,27 CH2F,27 or CF3.28-30 Other carboxylates used include benzoate,30,31 L(+)-mandelate,24 N-phenylanthranilate,32 and salicylate.33 Complexes with dienes other than 1,5COD such as norbornadiene (NBD),27,28,31,34,35 dicyclopentadiene,27 (ethylene)2,30 or (cyclooctene)2,30 have been attained, usually by varying the [(diene)Rh(µ-Cl)]2 starting material. Several comparable Ir compounds have been reported,27,30,36 including a patent that described the synthesis of “Ir(cyclooctadiene)(2-ethylhexanoate)” by addition of triethylammonium 2-ethylhexanoate to [(1,5-COD)Ir(µCl)]2.37 The previous literature describing syntheses of related Rh and Ir(1,5-COD)(µ-O2CR)-type compounds is summarized in Supporting Information, Table S1. (22) Mishra, S.; Daniele, S.; Hubert-Pfalzgraf, L. G. Chem. Soc. ReV 2007, 36, 1770–1787. (23) (a) Legzdins, P.; Mitchell, R. W.; Rempel, G. L.; Ruddick, J. D.; Wilkinson, G. J. Chem. Soc. A 1970, 3322–3326. (b) Hui, B. C.; Rempel, G. L. Chem. Commun. 1970, 1195–1196. (c) Hui, B. C. Y.; Teo, W. K.; Rempel, G. L. Inorg. Chem. 1973, 12, 757–762. (24) Nagy-Magos, Z.; Vastag, S.; Heil, B.; Marko´, L. J. Organomet. Chem. 1979, 171, 97–102. (25) Claver, C.; Ruiz, A.; Masdeu, A. M.; Ruiz, N. Inorg. Chim. Acta 1990, 175, 77–81. (26) Chatt, J.; Venanzi, L. M. J. Chem. Soc. 1957, 4735–4741. (27) Haszeldine, R. N.; Lunt, R. J.; Parish, R. V. J. Chem. Soc. A 1971, 23, 3696–3698. (28) Azbel, B. I.; Gol’Dshleger, N. F.; Khidekel, M. L.; Sokol, V. I.; PoraiKoshits, M. A. J. Mol. Catal. 1987, 40, 57–63. (29) Lahoz, F. J.; Martin, A.; Esteruelas, M. A.; Sola, E.; Serrano, J. L.; Oro, L. A. Organometallics 1991, 10, 1794–1799. (30) Werner, H.; Poelsma, S.; Schneider, M. E.; Windmu¨ller, B.; Barth, D. Chem. Ber. 1996, 129, 647–652. (31) Green, M.; Kuc, T. A. J. Chem. Soc., Dalton Trans. 1972, 832–839. (32) Trzeciak, A. M.; Zio´łkowski, J. J.; Lis, T.; Borowski, A. Polyhedron 1985, 4, 1677–1681.
1116 Inorganic Chemistry, Vol. 48, No. 3, 2009
remove the residual chloride salts and/or unreacted starting materials is important, however. Compounds 1 and 2 appear to be relatively air stable in crystalline form but not in solution.39 These points are discussed in greater detail in the Supporting Information for the interested reader. X-ray Crystal Structures of 1 and 2. The X-ray crystal structures of 1 and 2, with atomic numbering schemes and thermal ellipsoids at a 30% probability level, are shown in Figures 1 and 2, respectively.40 Compounds 1 and 2 both proved to be dimeric as expected,41 and are isomorphous. Each molecule has two 2-ethylhexanoate ligands bridging the two metal centers. Each transition metal center in 1 and 2 is four-coordinate, bound to one of the oxygen atoms from each 2-ethylhexanoate and to the two olefinic bonds from a single η2:2-1,5-COD. The centroids of the olefinic carbon (33) Mieczyn´ska, E.; Trzeciak, A. M.; Zio´łkowski, J. J.; Lis, T. Polyhedron 1994, 13, 655–658. (34) Reis, A. H., Jr.; Willi, C.; Siegel, S.; Tani, B. Inorg. Chem. 1979, 18, 1859–1863. (35) Chen, M. J.; Feder, H. M. Inorg. Chem. 1979, 18, 1864–1869. (36) Burk, M. J.; Crabtree, R. H. J. Am. Chem. Soc. 1987, 109, 8025– 8032. (37) Vaartstra, B. A. Metal Carboxylate Complexes for Formation of MetalContaining Films on Semiconductor Devices. U.S. Patent 5,695,815, Dec. 9, 1997. Characterization was accomplished therein by elemental analysis alone, which allowed for compositional determination but could not provide verification of the anticipated dimeric structure. Additionally, the synthesis was described therein with minimal detail, and no attempt to crystallize the compound was reported. (38) The synthesis, characterization, properties, and uses of metal 2-ethylhexanoates and their derivatives have been recently reviewed by Mishra et al.22 Metal 2-ethylhexanoates were described therein as “of great commercial importance” since they are used as driers in certain paints, lubricating agents, stabilizers for plastics, waterproofing agents, fuel additives, fungicides, in corrosion protection, and as desired precursors in materials science applications “to obtain nano-films, -composites, and -particles” by various deposition methods.22 For this wide variety of uses, they enjoy the “advantages of being inexpensive, air-stable, non-toxic as well as commercially available for a wide number of elements.”22 Metal carboxylates used in catalytic applications include Ni(2-ethylhexanoate)2 + AlEt3 for industrial polymer hydrogenation,1,2,22 Rh2(OAc)4 for olefin hydrogenation,23 and [(COD)Rh(O2CR)]2 + PR′3 for olefin hydrogenation,24 or hydroformylation.25 (39) It was claimed that 1 in the viscous liquid form was air stable,37 but this seems unlikely to be true, especially for air exposure at long time scales. Some air sensitivity is likely common for complexes of this type; for example, the related compound [(C2H4)2Rh(µ-O2CCH3)]2 was reported as “moderately air-sensitive.”30 (40) It has been noted that “single-crystal X-ray data on 2-ethylhexanoate derivatives remain scarce since the long carbon chain favors disorder and, as a result, the obtained crystals are often of poor quality.”22 For this reason, the relative ease with which 1 and 2 yielded X-ray diffraction-quality single crystals is notable. However, the crystal structures of 1 and 2 conform to the above statement in that disorder in the 2-ethylhexanoate alkyl chain consisting of C25C32 is pronounced. ¨ zkar, S.; Finke, R. G. J. Organomet. Chem. 2004, 689, 493–501. (41) O
Model Ziegler-Type Hydrogenation Catalyst Precursors Table 1. Selected Bond Lengths (Å) for 1 and 2
Figure 1. Single crystal X-ray diffraction structure and atomic numbering scheme for 1 with thermal ellipsoids at 30% probability. For the sake of clarity, hydrogen atoms are not shown. The large thermal ellipsoids for C28, C29 and C30 (here and in Figure 2) are as expected for such floppy alkyl chains.
Figure 2. Single crystal X-ray diffraction structure and atomic numbering scheme for 2 with thermal ellipsoids at 30% probability. For the sake of clarity, hydrogen atoms are not shown.
atoms reveals square planar geometries expected for d8 Ir(I) and Rh(I). However, in both complexes the square planes are imperfect: the metal atoms are displaced from their ligand planes by 0.1522 and 0.1328 Å for Ir1 and Ir2 respectively, and likewise, by 0.1293 and 0.1180 Å for Rh1 and Rh2, respectively, bringing the metal atoms closer together than they would otherwise be. Displacement from the ligand planes has been observed in similar complexes34 and has been interpreted as evidence of M-M interaction.42 The dihedral angle between the planes is 56.5° for 1 and 58.1° for 2, close to angles observed in similar complexes,42-44 some also with bridging carboxylates.34,45 The dihedral angle is influenced by the metal atom, the terminal ligands, and the type of bridging ligands for complexes of this type.46,47 (42) Fandos, R.; Herna´ndez, C.; Otero, A.; Rodrı´guez, A.; Ruiz, M. J. Organometallics 1999, 18, 2718–2723. (43) (a) Rodman, G. S.; Mann, K. R. Inorg. Chem. 1985, 24, 3507–3508. (b) Rodman, G. S.; Mann, K. R. Inorg. Chem. 1988, 27, 3338–3346. (44) Kanematsu, N.; Ebihara, M.; Kawamura, T. Inorg. Chim. Acta 1999, 292, 244–248. (45) Schnabel, R. C.; Roddick, D. M. Organometallics 1996, 15, 3550– 3555. (46) Sielisch, T.; Cowie, M. Organometallics 1988, 7, 707–714. (47) Aullo´n, G.; Ujaque, G.; Lledo´s, A.; Alvarez, S.; Alemany, P. Inorg. Chem. 1998, 37, 804–813.
bond
1, M ) Ir
2, M ) Rh
M1-C8 M1-C4 M1-O4 M1-C5 M1-O1 M1-C1 M2-O2 M2-C9 M2-C13 M2-C12 M2-C16 M2-O3 O1-C17 O2-C17 O3-C25 O4-C25
2.086(5) 2.086(3) 2.087(3) 2.097(4) 2.097(3) 2.103(4) 2.082(3) 2.084(4) 2.089(3) 2.093(4) 2.094(3) 2.101(3) 1.262(6) 1.261(5) 1.260(7) 1.254(7)
2.093(4) 2.088(3) 2.086(2) 2.104(3) 2.095(2) 2.105(3) 2.078(2) 2.087(3) 2.092(3) 2.098(3) 2.101(3) 2.100(2) 1.251(5) 1.266(4) 1.262(5) 1.254(5)
In addition, torsion angles about the M-M axes are 20.4° and 22.0° in 1 and 2, respectively, as determined using the [C1,C4,C5,C8] centroid-M1 and [C9,C12,C13,C16] centroidM2 vectors. This leads to a staggering of the two 1,5-COD ligands consistent with similar established structures.33,34,43,44,46 Electron counting for complexes 1 and 2 gives formal 16 electron metal centers with no formal M-M bonds. The M-M distances of structures related to 1 and 2, and what M-M interactions, if any, have been suggested for analogous complexes, can be found in Supporting Information, Table S2. The M-M distances in 1 and 2 of 3.2776(2) Å and 3.3390(4) Å,48 respectively, agree well with precedent for those compounds in which some extent of weak M-M interaction is believed likely.28,43b,49 Such axial M-M bonding interactions in dimeric, square planar, d8-d8 structures have been previously investigated,47,50 with it now understood that bonding occurs as a result of donor-acceptor interactions between filled dz2 and empty pz orbitals.47,50 Furthermore, bands in the UV-vis absorption spectra, at λmax ) 486 nm for 1, and λmax ) 422 nm for 2 (Supporting Information, Figure S5), correspond to bands assigned to metal-centered dσ*-pσ transitions in the UV-vis spectra of similar compounds.44,51,52 Selected bond lengths and angles with estimated standard deviations are given in Tables 1 and 2, respectively, for both compounds. The average M-O distance for 1 is 2.092(6) Å while for 2 it is 2.090(4) Å, close to expected values.30,34,45 The individual Ir-O distances in 1 are 2.097(3) and 2.082(3) Å for one 2-ethylhexanoate ligand and 2.087(3) and 2.101(3) Å for the other, each bound to Ir1 and Ir2, respectively. Likewise, the individual Rh-O distances in 2 are 2.095(2) and 2.078(2) Å for one 2-ethylhexanoate ligand, and 2.086(2) and 2.100(2) Å for the other, for Rh1 and Rh2, respectively. These distances indicate non-symmetrical bridging of the 2-ethylhexanoate ligands. Each metal center and each 2-eth(48) The formally non-bonding M-M distances and estimated standard deviations cited were calculated by imposing a M-M bond, then performing an additional least-squares refinement. (49) Coleman, A. W.; Eadie, D. T.; Stobart, S. R. J. Am. Chem. Soc. 1982, 104, 922–923. (50) Aullo´n, G.; Alvarez, S. Inorg. Chem. 1996, 35, 3137–3144. (51) Marshall, J. L.; Stobart, S. R.; Gray, H. B. J. Am. Chem. Soc. 1984, 106, 3027–3029. (52) Rodman, G. S.; Daws, C. A.; Mann, K. R. Inorg. Chem. 1998, 27, 3347–3353.
Inorganic Chemistry, Vol. 48, No. 3, 2009
1117
Alley et al. Table 2. Selected Bond Angles (deg) for 1 and 2 bond
1, M ) Ir
2, M ) Rh
C8-M1-C4 C8-M1-O4 C4-M1-O4 C8-M1-C5 O4-M1-C5 C8-M1-O1 C4-M1-O1 O4-M1-O1 C5-M1-O1 C4-M1-C1 O4-M1-C1 C5-M1-C1 O1-M1-C1 O2-M2-C9 O2-M2-C13 C9-M2-C13 O2-M2-C12 C9-M2-C12 O2-M2-C16 C13-M2-C16 C12-M2-C16 O2-M2-O3 C9-M2-O3 C13-M2-O3 C12-M2-O3 C16-M2-O3 C17-O1-M1 C17-O2-M2 C25-O3-M2 C25-O4-M1 O2-C17-O1 O2-C17-C18 O1-C17-C18 O4-C25-O3 O4-C25-C26 O3-C25-C26
99.17(17) 86.00(17) 166.32(14) 82.35(18) 153.88(15) 149.41(16) 91.47(13) 90.31(14) 87.98(15) 82.15(15) 94.26(16) 91.51(17) 170.54(14) 85.41(14) 166.05(15) 99.28(15) 154.13(15) 82.56(15) 93.49(13) 82.22(14) 91.69(15) 89.89(13) 151.15(15) 92.07(14) 89.72(14) 169.17(15) 130.9(3) 126.3(3) 128.7(3) 126.9(3) 125.6(4) 117.2(4) 117.2(4) 125.9(4) 118.6(6) 115.5(6)
99.08(14) 86.09(13) 164.91(12) 82.55(15) 155.97(13) 150.61(13) 90.91(11) 91.36(11) 88.31(12) 82.42(13) 93.18(13) 91.18(14) 169.94(12) 85.51(12) 165.32(12) 99.02(12) 155.52(12) 82.60(13) 92.77(11) 82.44(12) 91.30(12) 90.55(10) 152.13(12) 91.72(11) 90.06(11) 168.87(12) 131.9(2) 126.5(2) 128.7(2) 127.8(3) 126.2(3) 116.5(4) 117.3(3) 125.8(3) 117.0(4) 117.2(4)
ylhexanoate has one M-O bond longer than the other, with an average difference of 0.015(6) Å in 1 and 0.015(4) Å in 2. The average M-C distance is 2.09(1) in 1 and 2.096(9) Å in 2. The average olefinic CdC bond lengths are 1.42(1) and 1.398(9) Å in 1 and 2, respectively. The marginally longer CdC bond in 1 could be interpreted as the result of the more electron rich Ir(I) engaging in a greater degree of π-back-bonding than its Rh(I) counterpart. These values and interpretations are consistent with expectations based on similar structures.33,34,43,46 The R carbon of 2-ethylhexanoate is stereogenic. Complexes 1 and 2 crystallize with (R,R) and (S,S) enantiomers in a unit cell (centrosymmetric space group P21/n, Z ) 4). The possibility exists, therefore, to prepare chiral analogues of 1 and 2 (i.e., using enantiopure 2-ethylhexanoate, or analogous chiral ligands) to effect asymmetric catalysis. Demonstration of Catalytic Hydrogenation Activity. It is important to demonstrate that 1 and 2 form Ziegler-type hydrogenation catalysts. This was accomplished by the addition of a cyclohexane solution of AlEt3 to a cyclohexane solution of 1 or (separately) 2, at an Al/M ) 1 ratio; the result is an immediate change, from the orange solution of 1 to tawny yellow, or the yellow solution of 2 to clear brown. Representative cyclohexene hydrogenations are shown in Figures 3 and 4 for 1 and 2, respectively. Active H2 uptake begins immediately in both cases, suggesting that the products of the reaction between AlEt3 and 1 or 2 have preformed actiVe Ziegler-type hydrogenation catalysts (i.e.,
1118 Inorganic Chemistry, Vol. 48, No. 3, 2009
Figure 3. Two representative runs for the hydrogenation of cyclohexene using an Ir Ziegler-type hydrogenation catalyst formed upon combination of 1 and AlEt3, Al/Ir ) 1. Experimental conditions for all hydrogenations were solvent ) cyclohexane, temp. ) 22.0 °C, catalyst concentration ) 1.2 mM in [Metal], initial cyclohexene concentration ) 1.65 M, and stirring ) 1000 rpm. For clarity, only every fifth data point collected of run “a” is displayed. The absence of an induction period shows that either an active catalyst was present from the start of the hydrogenation, or conceivably formed essentially immediately once H2 was added. The change of the solution during the reaction to darker brown, and the subtle changes to the slope of these curves, suggest that further catalyst evolution is taking place during the hydrogenation.17
Figure 4. Two representative runs for the hydrogenation of cyclohexene using a Rh Ziegler-type hydrogenation catalyst formed upon combination 2 and AlEt3, Al/Rh ) 1. Again, the absence of an induction period demonstrates that either an active catalyst is present from the start of the hydrogenation, or forms essentially immediately once H2 was added. The catalyst formed using Rh 2 is >30 fold more active than its Ir 1 counterpart. Because of this high activity, the reproducibility of the two hydrogenation curves shown here is sensitive to the reproducibility of the H2 gas purge cycle, the start of data acquisition, and the precise stirring speed, as would be characteristic of reactions influenced by H2 gas-to-solution MTL.57
that the active catalysts are formed, at least in large part, from the addition of AlEt3, rather than being solely reduced or activated solely by H2). Catalyst formation from these systems is being optimized,17 according to variables known to be important to Ziegler-type hydrogenation catalysts,53-56 work that will be reported in due course. The linear versus normally expected exponential shape of both the curves in Figures 3 and 4 is of interest. The Rh catalyst made from 2 is significantly more active for cyclohexene hydrogenation than the Ir catalyst made using 1 (>30 times according to linear fits to the portions of the curve during which pressure loss was occurring, Supporting Information, Figure S10 and Table S3. The rates of H2
Model Ziegler-Type Hydrogenation Catalyst Precursors
pressure loss in Figure 4 are close to the H2 gas-to-solution MTL for the apparatus and conditions used,17 which accounts for the observed linearity.57 However, since the apparatus and conditions used are identical, the near-linearity of the >30-fold slower hydrogenation curves given by the catalyst made from 1, shown in Figure 3, cannot also be explained by MTL effects. Close inspection of the hydrogenation curves for 1 reveals subtle changes in the rate of H2 uptake during the hydrogenation. An additional point is that after all the cyclohexene had been consumed, the catalyst solutions appeared unchanged in the case of 2 (4-7 min), but had turned darker brown in the case of 1 (3.8 h). These observations imply that further catalyst development (i.e., in either consecutive or parallel reactions) is taking place as the cyclohexene hydrogenation proceeds in Figure 3. Detailed characterization studies using EXAFS, electron microscopy, and other methods are in progress,17 now that 1 and 2 are available and shown to yield active Ziegler-type hydrogenation catalysts. Control Experiments to Determine if AlEt3 is Required to Form Active Catalysts. Since a fundamental characteristic of Ziegler-type hydrogenation catalysts is their “activation” by an organometallic cocatalyst, frequently AlR3,1,2 control experiments were performed to check the necessity of AlEt3 in forming active hydrogenation catalysts from 1 and 2. Hydrogenations were performed identically to those in Figures 3 and 4, but leaving out the AlEt3 while substituting an equal volume of cyclohexane for the volume introduced by the AlEt3 solution. The results, Supporting Information, Figure S9, reveal that active, though poorly stabilized, nanoparticle and/or bulk metal catalysts are formed following an induction period. These control experiments confirm the need for, and value of, AlEt3 in producing longliVed hydrogenation catalysts from 1 and 2. Conclusions The synthesis, unequivocal characterization, and utility as precursors of Zieger-type hydrogenation catalysts of [(1,5COD)Ir(µ-O2C8H15)]2 (1) and [(1,5-COD)Rh(µ-O2C8H15)]2 (2) (COD ) cyclohexadiene, O2C8H15 ) 2-ethylhexanoate) have been described herein. Compositional and structural characterization of these compounds was carried out by elemental analysis, FAB-MS, single crystal X-ray diffractometry, FT-IR, UV-vis, and NMR spectroscopy. X-ray crystallography showed that 1 and 2 are isostructural. The complexes are carboxylate-bridged dimers with terminal η2:21,5-COD ligands. They have a bent geometry with dihedral (53) Alley, W. M.; Kayiran, I.; Johnson, K.; Finke, R. G. Ziegler-type Hydrogenation Catalysts made from Group 8-10 Transition Metal Precatalysts and AlR3 Cocatalysts: A Critical Review of the Literature. J. Mol. Catal. A: Chem., 2009, submitted for publication. (54) Sˇabata, S.; Hetflejsˇ, J. J. Appl. Polym. Sci. 2002, 85, 1185–1193. (55) Reguli, J.; Stasˇko, A. Chem. Pap. 1987, 41, 299–310. (56) Coolbaugh, T. S.; Loveless, F. C.; Mathews, D. N. Method of Synthesizing a Selective Olefin Hydrogenation Catalyst. European Patent Application 91300316.6, Jan 16, 1991. (57) Aiken, J. D., III; Finke, R. G. J. Am. Chem. Soc. 1998, 120, 9545– 9554. (58) Shriver, D. F.; Drezdzon, M. A. The Manipulation of Air-SensitiVe Compounds, 2nd ed.; John Wiley and Sons: New York, 1986.
angles of 56.5° and 58.1° for 1 and 2, respectively, between the square ligand planes of the 4-coordinate M(I) centers. Formally non-bonding M-M distances of 3.2776(2) and 3.3390(4) Å for 1 and 2, respectively, likely entail weak M-M bonding interactions.28,42,43b,44,47,49,50 Both 1 and 2 serve as effective precatalysts that, when combined with AlEt3, form highly active Ziegler-type hydrogenation catalysts. The Rh catalyst formed from 2 is g30-fold more active than the Ir catalyst prepared from 1. This suggests that it will prove useful to have both the Ir and Rh systems for comparison to industrial catalysts, commonly made from Ni or Co.1,2 Control hydrogenations performed without AlEt3 contained induction periods showing that 1 and 2 display a key feature of Ziegler-type hydrogenation catalysts by requiring “activation” by a cocatalyst, in this case AlEt3. Publication of our work on the characterization of the catalysts formed from 1 and 2 is forthcoming.17 That work teaches that the use of well-defined, model second and third row, precatalysts 1 and 2 has been instrumental in elucidating further insights into Ziegler-type hydrogenation catalysts,17 a ∼40 year old problem left unanswered by studies using less well-defined, first row industrial precursors. Experimental Section Procedures, Materials, and Instrumentation. Unless indicated otherwise, all manipulations were performed either under N2 in a Vacuum Atmospheres drybox or using air-free techniques on a Schlenk line. Oxygen levels were continuously maintained in the drybox at e5 ppm, monitored by a Vacuum Atmospheres O2-level monitor. All glassware was oven-dried at 160 °C overnight before use and cooled either under vacuum or under N2 in the drybox. Unless noted otherwise, all solvents, compounds, and other materials below were stored in the drybox. Cyclohexane was used as received (Sigma-Aldrich 99.5%, anhydrous: water 20% abs. (cm-1): 3030-2760m, 1560s, 1457s, 1419s, 1321s. The procedure was very similar for the synthesis of 2. Differences were that a solution of Na(2-ethylhexanoate) was used and the resulting crude product solution was filtered before the air-free hexane/H2O extraction. Crystals of 2 were obtained in an 86% yield. A recrystallization of a portion of the product 2 was required to produce larger, light orange, irregularly shaped single crystals suitable for X-ray diffractometry. Anal. Calcd for C32H54O4Rh2 (mol. wt. 708.57 g/mol): C, 54.24; H, 7.68; O, 9.0%; Na, 0 ppm. Found: C, 54.28; H, 7.87; O, 8.4%; Na 15% rel. intensity (m/z, rel. intensity, (estimated ion)+): 708.4, 65%, ([(1,5-COD)Rh(O2C8H15)]2)+; 565.0, 87%, ((1,5-COD)2Rh2(O2C8H15))+; 455.0, 55%, (C16H25O2Rh2)+; 415.0, 19%, ((C11H13O4Rh2)+; 353.1, base peak, ((1,5-COD)Rh(O2C8H15)-1)+; 309.9, 24%, (C13H19O2Rh)+; 211.0, 67%, ((1,5COD)Rh)+; 147.1, 16%, (Rh(O2C))+. 1H NMR in C6D6, (δ in ppm, multiplicity, no. of H): 4.30-4.40, m, 4; 4.18-4.30, m, 4; 2.65-2.82, m, 4; 2.15-2.38, m, 6; 1.20-1.72, m, 24; 0.85-0.98, m, 12. 13C NMR in C6D6, (δ in ppm): 187.8s, 80.8d, 74.0d, 50.9s,
1120 Inorganic Chemistry, Vol. 48, No. 3, 2009
Table 3. Summary of Crystallographic Data and Refinement for Compounds 1 and 2 compound
1
C32H54Ir2O4 887.15 100(1) 0.71073 P21/n 15.7748(5) 9.8962(3) 20.8847(7) 108.408(2) 3093.50(17) 1.905 4 8.629 R1 ) 0.0443, wR2 ) 0.0827 a R1 is for [I > 2σ(I)], wR2 is for all data: R1 ) wR2 ) [∑w(|Fo| - |Fc|)2/∑w|Fo|2]1/2. chemical formula formula weight T (K) λ (Å) space group a (Å) b (Å) c (Å) β (deg) V (Å3) Fcalcd (mg/m3) Z µ (mm-1) final R indicesa
2 C32H54O4Rh2 708.57 100(1) 0.71073 P21/n 15.7608(4) 9.9032(3) 20.8259(5) 108.527(1) 3082.09(14) 1.527 4 1.105 R1 ) 0.0504, wR2 ) 0.1260 ∑||Fo| - |Fc||/∑|Fo|,
33.4s, 31.9s, 31.0s, 30.6s, 27.1s, 23.6s, 14.7s, 12.9s. IR bands >20% abs. (cm-1): 3040-2760m, 1567s, 1462s, 1415s, 1324s, 952s, 817s. X-ray Crystallographic Structure Determination and Refinement. X-ray diffraction data were collected on a Bruker Kappa APEXII X-ray diffractometer equipped with a beam monochromator. Corrections applied were Lorentz, polarization, and absorption (SADABS).59 Data collection and cell refinement were accomplished using Bruker SMART software, and data reduction using Bruker SAINT. Both structures were solved by direct methods. Structure solution, structure refinement, and figure preparation were achieved using Bruker SHELXTL.60 Refinements were accomplished by full-matrix weighted least-squares on F2 for all reflections. Anisotropic displacement parameters were used for refinement of all non-hydrogen atoms. Hydrogen atoms were included at idealized positions in structure factor calculations. Experimental details for crystal data and structural refinement of 1 and 2 are displayed in Table 3. Precatalyst and Catalyst Solution Preparation. The Ir and Rh precatalysts 1 and 2 were used in catalyst preparation by first making a stock solution of each in cyclohexane. Specifically, a stock solution of 1, 7.20 mM in [Ir], was prepared by weighing out 0.0430 ( 0.0001 g (0.0485 mmol) of crystalline 1 and dissolving with 13.46 ( 0.01 mL of cyclohexane. A stock solution of 2, 7.20 mM in [Rh] was prepared by weighing out 0.04337 ( 0.0001 g (0.0612 mmol) of crystalline 2, and dissolving in 17.00 ( 0.01 mL of cyclohexane. AlEt3 was used as a 36.0 mM stock solution in cyclohexane prepared by adding 50-70 mL of cyclohexane to a 100 mL volumetric flask, followed by 0.53 ( 0.01 mL of neat AlEt3 measured out by syringe, and then diluting to the mark (this included a 7% correction factor to take into account the 93% purity of the AlEt3, even though the impurities are primarily other aluminum alkyls). Catalyst solutions were prepared individually, from the stock precatalyst and AlEt3 cocatalyst solutions, in new 22 × 175 mm Pyrex borosilicate culture tubes containing new 5/8 × 5/16 in. Teflon-coated magnetic stirbars. Both culture tube and stirbar had been cleaned by rinsing three times with nanopure water prior to drying at 160 °C overnight and cooled either under vacuum or under N2 in the drybox. Specifically, and using 1 as an example, a catalyst solution 1.2 mM in [Ir] was prepared individually, and in the drybox, by first adding 0.50 ( 0.01 mL of 7.2 mM 1 to a culture (59) Sheldrick, G. M. SADABS, a program for Siemens Area Detection Absorption Correction; Bruker AXS, Inc.: Madison, WI, 2000. (60) Sheldrick, G. M. Acta Cryst. A 2008, 64, 112–122.
Model Ziegler-Type Hydrogenation Catalyst Precursors tube, followed by 1.90 ( 0.02 mL of cyclohexane (Sigma-Aldrich 99.5%, anhydrous: water
